← back to catalog · registered 2026-11-21 15:04

lyber-ai/Hy4-flash-UNCENSORED-NVFP4

Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/lyber-ai%2FHy4-flash-UNCENSORED-NVFP4"
Response includes
  • classification m-uncensored
  • files 161
  • author_summary 8 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M-U
Primary method

Uncensored (method unknown)

No other method signals detected in this model.
Confidence
LOW
Why this label 3 signals
Weak or ambiguous signals. Best guess based on catalog patterns; treat as tentative and check the evidence below.
  • 'uncensored' in name/tags but no 'abliterated' marker
  • method not identifiable from author declaration alone
  • may be DPO fine-tune, prompt engineering, or unknown technique
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · 30-day
0
Likes
1
Model age
today
created 2026-11-21
Downloads over time
Now 0 → from 0 ↑ 0%
0011 0 on Nov 21 0 on Nov 22 Nov
Nov 21 → Nov 22 · 2 snapshots · spans 1 day

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Variants by this author 3 formats · 0 downloads combined

The same weights this author released in different packaging. Pick the format that matches your runtime.

Metadata

License
other
Languages
en zh es fr ja ko ru
Tags
transformers safetensors hy_v4_flash text-generation tencent hunyuan hy4 hy4-flash nvfp4 modelopt w4a16 uncensored

Related

Total size
204 GB
Files
161
Quantizations
1
Registered
2026-11-21 15:04
Last updated on HF
2026-11-21 22:25

Files by quantization

Auxiliary files 161 files 204 GB
model-00001-of-00148.safetensors 29.5 GB ******** download
model-00002-of-00148.safetensors 29.5 GB ******** download
model-00003-of-00148.safetensors 29.5 GB ******** download
model-00004-of-00148.safetensors 29.5 GB ******** download
model-00148-of-00148.safetensors 1.58 GB ******** download
model-00005-of-00148.safetensors 1.27 GB ******** download
model-00006-of-00148.safetensors 1.27 GB ******** download
model-00008-of-00148.safetensors 1.27 GB ******** download
model-00009-of-00148.safetensors 1.27 GB ******** download
model-00011-of-00148.safetensors 1.27 GB ******** download
model-00015-of-00148.safetensors 1.27 GB ******** download
model-00017-of-00148.safetensors 1.27 GB ******** download
model-00018-of-00148.safetensors 1.27 GB ******** download
model-00020-of-00148.safetensors 1.27 GB ******** download
model-00021-of-00148.safetensors 1.27 GB ******** download
model-00024-of-00148.safetensors 1.27 GB ******** download
model-00027-of-00148.safetensors 1.27 GB ******** download
model-00029-of-00148.safetensors 1.27 GB ******** download
model-00030-of-00148.safetensors 1.27 GB ******** download
model-00032-of-00148.safetensors 1.27 GB ******** download
model-00033-of-00148.safetensors 1.27 GB ******** download
model-00036-of-00148.safetensors 1.27 GB ******** download
model-00038-of-00148.safetensors 1.27 GB ******** download
model-00039-of-00148.safetensors 1.27 GB ******** download
model-00041-of-00148.safetensors 1.27 GB ******** download
model-00044-of-00148.safetensors 1.27 GB ******** download
model-00045-of-00148.safetensors 1.27 GB ******** download
model-00047-of-00148.safetensors 1.27 GB ******** download
model-00052-of-00148.safetensors 1.27 GB ******** download
model-00062-of-00148.safetensors 1.27 GB ******** download
model-00067-of-00148.safetensors 1.27 GB ******** download
model-00073-of-00148.safetensors 1.27 GB ******** download
model-00085-of-00148.safetensors 1.27 GB ******** download
model-00092-of-00148.safetensors 1.27 GB ******** download
model-00107-of-00148.safetensors 1.27 GB ******** download
model-00128-of-00148.safetensors 1.27 GB ******** download
model-00138-of-00148.safetensors 1.27 GB ******** download
model-00012-of-00148.safetensors 1.27 GB ******** download
model-00014-of-00148.safetensors 1.27 GB ******** download
model-00023-of-00148.safetensors 1.27 GB ******** download
model-00026-of-00148.safetensors 1.27 GB ******** download
model-00035-of-00148.safetensors 1.27 GB ******** download
model-00042-of-00148.safetensors 1.27 GB ******** download
model-00078-of-00148.safetensors 1.27 GB ******** download
model-00099-of-00148.safetensors 1.27 GB ******** download
model-00117-of-00148.safetensors 1.27 GB ******** download
model-00065-of-00148.safetensors 795 MB ******** download
model-00095-of-00148.safetensors 725 MB ******** download
model-00007-of-00148.safetensors 648 MB ******** download
model-00010-of-00148.safetensors 648 MB ******** download
model-00013-of-00148.safetensors 648 MB ******** download
model-00016-of-00148.safetensors 648 MB ******** download
model-00019-of-00148.safetensors 648 MB ******** download
model-00022-of-00148.safetensors 648 MB ******** download
model-00025-of-00148.safetensors 648 MB ******** download
model-00028-of-00148.safetensors 648 MB ******** download
model-00031-of-00148.safetensors 648 MB ******** download
model-00034-of-00148.safetensors 648 MB ******** download
model-00037-of-00148.safetensors 648 MB ******** download
model-00040-of-00148.safetensors 648 MB ******** download
model-00046-of-00148.safetensors 648 MB ******** download
model-00048-of-00148.safetensors 648 MB ******** download
model-00049-of-00148.safetensors 648 MB ******** download
model-00053-of-00148.safetensors 648 MB ******** download
model-00054-of-00148.safetensors 648 MB ******** download
model-00063-of-00148.safetensors 648 MB ******** download
model-00064-of-00148.safetensors 648 MB ******** download
model-00068-of-00148.safetensors 648 MB ******** download
model-00069-of-00148.safetensors 648 MB ******** download
model-00074-of-00148.safetensors 648 MB ******** download
model-00079-of-00148.safetensors 648 MB ******** download
model-00087-of-00148.safetensors 648 MB ******** download
model-00093-of-00148.safetensors 648 MB ******** download
model-00094-of-00148.safetensors 648 MB ******** download
model-00108-of-00148.safetensors 648 MB ******** download
model-00109-of-00148.safetensors 648 MB ******** download
model-00118-of-00148.safetensors 648 MB ******** download
model-00119-of-00148.safetensors 648 MB ******** download
model-00129-of-00148.safetensors 648 MB ******** download
model-00130-of-00148.safetensors 648 MB ******** download
model-00043-of-00148.safetensors 648 MB ******** download
model-00057-of-00148.safetensors 648 MB ******** download
model-00058-of-00148.safetensors 648 MB ******** download
model-00080-of-00148.safetensors 648 MB ******** download
model-00086-of-00148.safetensors 648 MB ******** download
model-00100-of-00148.safetensors 648 MB ******** download
model-00101-of-00148.safetensors 648 MB ******** download
model-00139-of-00148.safetensors 648 MB ******** download
model-00140-of-00148.safetensors 648 MB ******** download
model-00055-of-00148.safetensors 632 MB ******** download
model-00059-of-00148.safetensors 302 MB ******** download
model-00050-of-00148.safetensors 288 MB ******** download
model-00070-of-00148.safetensors 284 MB ******** download
model-00075-of-00148.safetensors 279 MB ******** download
model-00081-of-00148.safetensors 265 MB ******** download
model-00088-of-00148.safetensors 262 MB ******** download
model-00120-of-00148.safetensors 254 MB ******** download
model-00102-of-00148.safetensors 217 MB ******** download
model-00143-of-00148.safetensors 176 MB ******** download
model-00112-of-00148.safetensors 156 MB ******** download
model-00103-of-00148.safetensors 140 MB ******** download
model-00122-of-00148.safetensors 131 MB ******** download
model-00110-of-00148.safetensors 120 MB ******** download
model-00133-of-00148.safetensors 116 MB ******** download
model-00141-of-00148.safetensors 115 MB ******** download
model-00131-of-00148.safetensors 105 MB ******** download
model-00097-of-00148.safetensors 104 MB ******** download
model-00121-of-00148.safetensors 86.9 MB ******** download
model-00113-of-00148.safetensors 81.2 MB ******** download
model-00144-of-00148.safetensors 76.8 MB ******** download
model-00096-of-00148.safetensors 69.1 MB ******** download
model-00123-of-00148.safetensors 67.2 MB ******** download
model-00132-of-00148.safetensors 65.3 MB ******** download
model-00142-of-00148.safetensors 65.3 MB ******** download
model-00089-of-00148.safetensors 61.8 MB ******** download
model-00111-of-00148.safetensors 57.4 MB ******** download
model-00124-of-00148.safetensors 45.9 MB ******** download
model-00114-of-00148.safetensors 42.4 MB ******** download
model-00134-of-00148.safetensors 36.0 MB ******** download
model-00145-of-00148.safetensors 29.3 MB ******** download
model-00135-of-00148.safetensors 27.0 MB ******** download
model-00071-of-00148.safetensors 24.0 MB ******** download
model-00104-of-00148.safetensors 21.4 MB ******** download
model-00076-of-00148.safetensors 18.9 MB ******** download
model-00136-of-00148.safetensors 18.9 MB ******** download
model-00082-of-00148.safetensors 13.5 MB ******** download
model-00115-of-00148.safetensors 9.00 MB ******** download
model-00125-of-00148.safetensors 9.00 MB ******** download
model-00077-of-00148.safetensors 9.00 MB ******** download
model-00091-of-00148.safetensors 9.00 MB ******** download
model-00137-of-00148.safetensors 9.00 MB ******** download
model-00083-of-00148.safetensors 7.88 MB ******** download
model-00105-of-00148.safetensors 7.88 MB ******** download
model-00090-of-00148.safetensors 6.65 MB ******** download
model-00056-of-00148.safetensors 5.37 MB ******** download
model-00098-of-00148.safetensors 5.06 MB ******** download
model-00146-of-00148.safetensors 5.06 MB ******** download
model-00066-of-00148.safetensors 4.50 MB ******** download
model-00060-of-00148.safetensors 3.94 MB ******** download
model-00084-of-00148.safetensors 3.94 MB ******** download
model-00126-of-00148.safetensors 3.94 MB ******** download
model-00116-of-00148.safetensors 3.94 MB ******** download
model-00051-of-00148.safetensors 2.53 MB ******** download
model-00106-of-00148.safetensors 1.69 MB ******** download
model-00147-of-00148.safetensors 1.33 MB ******** download
model-00061-of-00148.safetensors 972 KB ******** download
model-00072-of-00148.safetensors 972 KB ******** download
model-00127-of-00148.safetensors 972 KB ******** download
model.safetensors.index.json 23.2 MB ******** download
tokenizer.json 8.93 MB ff6aaf29 download
hf_quant_config.json 42.7 KB ecea9d4b download
chat_template.jinja 16.8 KB 974eeefa download
LICENSE 13.6 KB 2be6f885 download
dealign_mascot.png 10.9 KB da3bf39a download
config.json 10.5 KB 53b66f45 download
README.md 8.72 KB 750661d6 download
processor_config.json 1.60 KB 6c74c1c7 download
.gitattributes 1.55 KB 495a7175 download
lyber_icon.png 1.28 KB f8b116fd download
tokenizer_config.json 1.25 KB 59ac7739 download
generation_config.json 189 B aa393606 download

README current version from Hugging Face


language:

  • en
  • zh
  • es
  • fr
  • ja
  • ko
  • ru
    license: other
    license_name: tencent-hunyuan
    license_link: https://huggingface.co/tencent/Hy4-flash/blob/main/LICENSE
    base_model: tencent/Hy4-flash
    base_model_relation: finetune
    library_name: transformers
    pipeline_tag: text-generation
    thumbnail: dealign_mascot.png
    extra_gated_prompt: |
    By requesting access you confirm you are 18+, will use the model solely in
    compliance with applicable law in your jurisdiction, and accept full responsibility
    for all outputs. Access is granted at the sole discretion of the repository owner.
    extra_gated_fields:
    Name: text
    Affiliation / organization: text
    Country: country
    Primary use case: text
    I confirm I am 18 or older: checkbox
    I accept full responsibility for outputs: checkbox
    extra_gated_heading: "Request access — manual approval required"
    extra_gated_description: "This model is reviewed manually. Fill in the fields below."
    extra_gated_button_content: "Request access"
    tags:
  • tencent
  • hunyuan
  • hy4
  • hy4-flash
  • nvfp4
  • modelopt
  • w4a16
  • uncensored
  • reasoning
  • tool-calling
  • moe

   

A Lyber AI × dealign.ai release

Twitter — @dealignai
Lyber AI

[!NOTE]
Please credit if you build on this. If you redistribute, quantize, fine-tune, serve, benchmark, or write about this model, a line like
"Based on lyber-ai/Hy4-flash-UNCENSORED-NVFP4"
in your model card or write-up means a lot. Thank you.

[!IMPORTANT]
NVFP4 (modelopt) weight-only · group_size=16 · F8_E4M3 per-group scales ·
F32 global weight_scale_2 · dragonfly / vision / engram / HC paths kept BF16/F32.
219 GB disk (vs 411 GB BF16 / 267 GB FP8). Produced via direct-shards quantization
in Nvidia modelopt NVFP4 format.

[!WARNING]
Serving note (as of 2026-11-21). Public vllm/vllm-openai:hy4-flash (2026-11-20),
lmsysorg/sglang:hy4 (2026-11-20), and lmsysorg/sglang:nightly-cu134-20261121
images do not yet fully support serving NVFP4 Hy4-flash. In sglang, --quantization modelopt_fp4 wires up the MoE-experts path (routed w13 / w2), but the
non-MoE Linears used by Hy4-flash — attention q_b_proj / kv_b_proj / fused
qkv_a_proj_with_mqa, the dense MLP at layer 0, and (before patching) the
shared_experts block — are instantiated without NVFP4 parameter registration,
so their weight_scale / weight_scale_2 destinations don't exist at load. The
artifact is structurally identical to
atakans/Hy4-flash-NVFP4
(same tensor naming, shapes, and dtypes), so it will serve the moment upstream
SGLang / vLLM land NVFP4 parameter registration for Hy4-flash's full Linear set
.
Today, use the FP8 variant lyber-ai/Hy4-flash-UNCENSORED-FP8
for direct SGLang / vLLM serving, or run these NVFP4 shards through TensorRT-LLM
(Nvidia's native modelopt NVFP4 serve stack).

Hy4-flash — UNCENSORED (NVFP4)

Uncensored release · NVFP4 modelopt · 128K context · Extended thinking · Tools


What this is

This is tencent/Hy4-flash (Hunyuan V4
Flash) with refusal behavior removed, then quantized to NVFP4 in Nvidia modelopt
format. All upstream capabilities — extended thinking with reasoning_effort tiers,
tool calling, DSA attention, dragonfly MTP fusion, 128K+ context — are preserved.

Quantization details

  • Format: modelopt NVFP4, hf_quant_config.json at root (quant_algo: NVFP4,
    group_size: 16), compatible with upstream Nvidia modelopt / TensorRT-LLM and
    (pending full support) vLLM --quantization modelopt_fp4 + SGLang
    --quantization modelopt_fp4.
  • FP4 e2m1 levels: {±0, ±0.5, ±1, ±1.5, ±2, ±3, ±4, ±6} (16 values).
  • Scaling (per weight tensor):
    • .weight : uint8, 2 FP4 values per byte, shape (out, in/2)
    • .weight_scale : F8_E4M3, per-group (16-wide) scale, shape (out, in/16)
    • .weight_scale_2 : F32 scalar (global amax renormalization so per-group scales stay inside F8_E4M3 range)
    • .input_scale : F32 scalar = 1.0 (weight-only NVFP4; placeholder for activation path)
  • Dequantization: W_fp32 = packed_fp4 × weight_scale.to(fp32) × weight_scale_2
  • MoE experts: per-expert, unfused (experts.{i}.gate_proj, experts.{i}.up_proj,
    experts.{i}.down_proj) matching upstream NVFP4 conventions; 512 experts × 34 MoE
    layers.
  • Excluded from quantization (kept BF16/F32): lm_head, all model.dfly_layers.*
    (dragonfly MTP fusion), vit.* (vision tower), engram_layer.* (Engram speculative
    memory), self_attn.compressed_indexer.*, hc_pre / hc_attn / hc_mlp hyperconnection
    parameters, learnable_sink_param, router gate.weight.
  • Shard count: 148. Total size: 219 GB.

Reference ancestry

  • Refusal behavior removed from the BF16 counterpart of tencent/Hy4-flash (same
    method and strengths used in the FP8 companion release
    lyber-ai/Hy4-flash-UNCENSORED-FP8).
  • NVFP4 format matches atakans/Hy4-flash-NVFP4
    (producer: modelopt 0.48.0, group_size 16, same tensor naming and dtypes).

FP8-counterpart benchmarks (same base crack, served as FP8)

These are the measured gates on the shipped FP8 variant, which is produced from the
same BF16-cracked base this NVFP4 quantizes. NVFP4-specific re-measurement will be
added here once serving support lands.

gate result
HB-320 full high COMPLY 306/320 (95.6%) (content 256 + reasoning-engaged 50)
HB-320 full high borderline 14/320 (4.4%)
HB-320 full no_think COMPLY 313/320 (97.8%)
MMLU-228 no_think 87.28% (base 88.16% → Δ −0.88 pp)
MMLU-228 high 91.67% (base 94.30% → Δ −2.63 pp)
harsh-MT preserved (3 seeds × 10 turns × 49K) 0 loops / 0 no_answer / mean_entropy ~0.30 nats
harsh-MT content (3 seeds × 10 turns × 49K) 0 loops / 0 no_answer / mean_entropy ~0.30 nats
30-turn stress preserved (2 seeds × 30 turns) 0 errors, 40K+ context
loop_check benign (72 × 2 modes) 4 loops no_think + 2 loops high (2.8% residual)

How to serve (when supported)

TensorRT-LLM (Nvidia modelopt native)

# Build engine from modelopt NVFP4 checkpoint
trtllm-build \
  --checkpoint_dir /path/to/Hy4-flash-UNCENSORED-NVFP4 \
  --output_dir /path/to/engine \
  --gemm_plugin fp4 --gpt_attention_plugin bfloat16 \
  --max_batch_size 32 --max_input_len 8192 --max_output_len 24576

vLLM (pending NVFP4 Hy4-flash support)

docker run -it --rm --gpus all --ipc=host --shm-size=64g \
  -v /path/to/model:/model \
  vllm/vllm-openai:latest \
    --model /model \
    --served-model-name hy4-flash \
    --tensor-parallel-size 8 \
    --max-model-len 131072 \
    --quantization modelopt_fp4 \
    --trust-remote-code \
    --host 0.0.0.0 --port 8000

SGLang (pending NVFP4 Hy4-flash support)

docker run -d --gpus all --ipc=host --shm-size=64g \
  -v /path/to/model:/model \
  -p 8000:8000 \
  --entrypoint python3 \
  lmsysorg/sglang:hy4 \
  -m sglang.launch_server \
    --model-path /model --served-model-name hy4-flash \
    --tp-size 8 --quantization modelopt_fp4 --port 8000

Hardware requirements (NVFP4)

GPU Setup Status
4× B300 SXM6 TP4 ✅ Fits (219 GB in 564 GB VRAM)
8× H200 141GB TP8, FP4 kernels ✅ Fits
8× H100 80GB TP8, FP4 kernels ✅ Fits
2× B300 TP2 ⚠️ Tight, context-window dependent

Reasoning + tools

chat_template_kwargs = {"reasoning_effort": "high"}  # or "no_think"

Responsible use

You are solely responsible for how you use this model and its outputs.

  • Do not use for generating illegal content in your jurisdiction
  • Do not use to harm, deceive, or exploit real individuals
  • Do not use to generate CSAM or non-consensual intimate content
  • Comply with all applicable laws and regulations

License

Inherits the Tencent Hy License, Version 1.0
from the base model. The modification does not change the license terms.

Credits

Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration